| Message ID | 20230505073813.1219175-1-perex@perex.cz |
|---|---|
| Headers | show |
| Series | rewrite snd_pcm_playback_silence() again | expand |
On Fri, 05 May 2023 09:38:08 +0200, Jaroslav Kysela wrote: > > This is a split of changes for both patches (the first with the missing > top-up mode + later fix with mixups) proposed by Oswald. The aim was to > pick only real code changes. > > Only the first two patches after revert fixes the current silencing issues. > The last two are just cleanups with the extra optimization in the last patch > moving the common code to a function. > > Cc: Oswald Buddenhagen <oswald.buddenhagen@gmx.de> > Cc: Jeff Chua <jeff.chua.linux@gmail.com> > > Jaroslav Kysela (5): > ALSA: pcm: Revert "ALSA: pcm: rewrite snd_pcm_playback_silence()" > ALSA: pcm: fix playback silence - use the actual new_hw_ptr for the > threshold mode > ALSA: pcm: fix playback silence - correct the incremental silencing > ALSA: pcm: playback silence - remove extra code > ALSA: pcm: playback silence - move silence variables updates to > separate function Thanks, this makes it much easier to digest the whole changes! I watch out for a while whether any objection comes up, then apply patches in today -- or if we didn't reach to consensus, I'll pick up only the first revert patch for 6.4-rc1, at least. Takashi
On Fri, May 05, 2023 at 11:31:16AM +0200, Takashi Iwai wrote: >On Fri, 05 May 2023 09:38:09 +0200, >Jaroslav Kysela wrote: >> Signed-off-by: Jaroslav Kysela <perex@perex.cz> > >Better to add Reported-by and the reference for the original thread >reporting the regression, > i'll post a slight rework of the series shortly, where i'll include this. > as well as the Fixes tag. > that seems pointless for a revert, as all the info is already included anyway, no? regards
On Fri, May 05, 2023 at 11:57:00AM +0200, Takashi Iwai wrote: >But, the difficulty of the code (even after this patch) is that the >filling behavior is completely different between the threshold and the >fill-up modes, and we still try to use the similar code. > what is slightly confusing is that we're kinda abusing silence_filled to mean silence_filled + real_samples. i can add a comment to that effect, but i don't think it's worth tearing apart the paths. regards
This is a split of changes for both patches (the first with the missing top-up mode + later fix with mixups) proposed by Oswald. The aim was to pick only real code changes. Only the first two patches after revert fixes the current silencing issues. The last two are just cleanups with the extra optimization in the last patch moving the common code to a function. Cc: Oswald Buddenhagen <oswald.buddenhagen@gmx.de> Cc: Jeff Chua <jeff.chua.linux@gmail.com> Jaroslav Kysela (5): ALSA: pcm: Revert "ALSA: pcm: rewrite snd_pcm_playback_silence()" ALSA: pcm: fix playback silence - use the actual new_hw_ptr for the threshold mode ALSA: pcm: fix playback silence - correct the incremental silencing ALSA: pcm: playback silence - remove extra code ALSA: pcm: playback silence - move silence variables updates to separate function sound/core/pcm_lib.c | 91 +++++++++++++++++++++++++---------------- sound/core/pcm_local.h | 3 +- sound/core/pcm_native.c | 6 +-- 3 files changed, 61 insertions(+), 39 deletions(-)